What Is Warehouse Slotting, and Should You Fix It Before Routing?

Warehouse slotting is deciding which storage location each product occupies. Good slotting puts frequently picked items near the packing area, and puts items that are often ordered together near each other. Poor slotting is the largest single cause of unnecessary walking in a picking operation.

It is one of three levers people use to reduce picking effort. The other two are batching, which is choosing which orders get picked together, and routing, which is the path a picker takes once the trip is set. Slotting is usually the highest-leverage of the three, for a reason that is easy to state and easy to miss: routing makes a trip shorter, and slotting removes the trip.

Usually, though, is doing real work in that sentence. Which of the three matters most in your building depends on the shape of your orders, and that is the first thing worth establishing. This page covers what slotting is, when it beats routing and when it does not, the arithmetic to run on your own numbers, and the things that have to be true before any of it is worth starting.

The three levers, and why they are not interchangeable

They act on different parts of the same problem, and it is worth being precise about which.

• Slotting decides where each item lives. It sets which locations an order will require somebody to visit. It is the only one of the three that changes the underlying geography of the work.
• Batching decides which orders get picked on the same trip. It spreads the fixed cost of walking across more lines. It does not change where anything is.
• Routing decides the sequence of stops within a trip that has already been defined. It takes the locations as given and finds a shorter path between them.

Read in that order they form a hierarchy. Slotting determines the set of stops. Batching determines how many orders share those stops. Routing determines the order of them. Each one operates inside the constraints the one above it has already set.

That is why the sequence of work matters. Optimising a route through a badly slotted warehouse produces the best possible path through an arrangement that should not exist. The optimisation is real, it is just bounded by a decision nobody revisited.

Why pickers walk as far as they do

The items on a typical order are stored in different places. Where each one was placed was usually decided once, when it first arrived, based on what space happened to be free that day. Nobody wrote the reasoning down because there was not much reasoning to write down.

That decision was made once. Every order since has been paying for it.

The walking is not caused by the pickers and it is not caused by the orders. It is the accumulated result of thousands of small, reasonable, undocumented placement decisions, none of which was wrong at the time.

In most picking operations, roughly half of the time goes on walking rather than picking. That figure traces to Frazelle’s work of 1996, which is the source most of the industry is quoting whenever it says travel dominates picking time. We have not found a methodology disclosed anywhere in the citation chain, and we would rather say that than repeat the number as though it were settled. Treat it as the reason to go and measure your own, not as a substitute for measuring it.

Slotting or routing first?

Slotting, in most operations. This is the single most useful thing on this page and it is the thing most operations get backwards.

Routing optimisation finds a shorter path between locations you have to visit. Slotting changes which locations you have to visit at all. Routing improves a trip. Slotting can eliminate it.

Consider an order with four lines on it. Routing takes the four locations those items live in and sequences them so the walk is as short as possible. That is real work and it is worth doing. But the four locations are a given. If two of those items were stored next to each other instead of at opposite ends of the building, the trip would be shorter than the best possible route through the current layout.

Routing works within the constraint. Slotting changes the constraint.

So why does routing get most of the attention? Because it is easier to change. Routing is a software setting. Slotting means physically moving pallets, relabelling locations, and retraining the people who know where things are. One of those can be done on a Tuesday afternoon and the other needs a plan, a budget and somebody senior to sign it off.

That is a real difference in difficulty and it deserves respect rather than dismissal. It is not, however, a reason to believe the easier lever is the bigger one.

The variable that actually decides it: lines per order

Here is the qualification that most slotting material leaves out, and any operator reading this will already have thought of it.

How much routing can possibly do for you is determined by how many stops a trip has. That comes from your order profile.

If your orders average one line, routing has nothing to sequence. There is one stop. The only question that exists is how far that one item is from the packing area, which is a slotting question and nothing else. In a single-line operation, slotting is not the biggest lever, it is very close to the only one, and batching is the other half of the conversation.

If your orders average twenty lines spread across the building, the picture changes. The trip now has real structure, sequencing it badly costs a great deal, and routing starts to earn its attention. Slotting still sets the geography those twenty stops sit in, so it has not stopped mattering. It has stopped being the only thing that matters.

So the honest version of the advice is this. Pull your average lines per order and the distribution behind the average, because the average alone will mislead you if you have a mix of small replenishment orders and large wholesale ones. The lower that number, the more completely slotting dominates. The higher it is, the more routing and batching deserve a seat at the table. In neither case does routing change which locations exist, which is why slotting is still the thing to look at first.

Slotting is two different jobs

Most material on this subject treats slotting as one activity. It is two, they use different data, and they can pull against each other.

• Velocity slotting ranks items by how often they are picked and puts the busiest ones closest to the packing area. It uses one number per item, it is simple to explain, and it is where nearly every slotting project starts.
• Affinity slotting looks at which items appear on the same order as each other and stores those near each other. It uses pairs rather than single items, and it is the one that determines how long a multi-line trip is.

The distinction matters because velocity slotting alone can make a multi-line operation worse. Rank ten thousand items by pick frequency, place them in that order outward from packing, and two items that ship together on forty orders a week can easily end up in different zones because their individual frequencies are different. Every one of those forty orders now crosses the building.

Velocity answers how far one pick is. Affinity answers how far a trip is. If your orders are single line, velocity is the whole job. The more lines your orders carry, the more affinity is doing the work, and the more a velocity-only analysis will disappoint you after the pallets have already been moved.

This is also the practical reason the order history matters more than the algorithm. Velocity can be computed from a simple count. Affinity only exists in the record of which lines appeared on the same order as each other, which means it can only be recovered from real order history and cannot be reconstructed from current stock or from an item master.

What the calculation actually looks like

Nobody should authorise moving pallets around a building on the strength of a principle. Here is the arithmetic, and the point of showing it is that you should run it on your own numbers rather than accept anyone else’s.

Start with the travel time you are actually carrying: pickers on the floor, multiplied by productive hours per shift, multiplied by the share of that time spent travelling.

Take an operation running 20 pickers over a 7.5 hour productive shift. If travel is around half the shift, that is 75 picker-hours of walking per shift. Not per week. Per shift.

Now the part where you supply the assumption rather than borrowing one. Suppose a re-slot removes one hour in every ten of that travel. That is 7.5 hours given back per shift.

Here is where most business cases go wrong. Seven and a half hours is not one extra person, and anyone who presents it that way is selling something. Three things stand between the recovered hours and the benefit.

• The work has to exist. Recovered travel time only converts into output if there are more lines waiting to be picked. In an operation that finishes its shift with time to spare, the saving shows up as idle time, not throughput.
• Every move creates replenishment work. This one is large enough that it gets its own section below.
• Moving costs labour once. The re-slot itself takes people and time, and it happens before any of the benefit arrives.

Do the netting and then express the result in the number your operation is already measured on. Not distance, and not travel time, neither of which appears in the WERC DC Measures benchmark. Lines picked and shipped per person hour, which does.

In the example above, 20 pickers over a 7.5 hour shift is 150 person-hours. An operation picking at 100 lines per person hour ships 15,000 lines in that shift. Recover 7.5 hours of travel, keep the headcount the same, and have the demand to fill it, and you ship roughly 750 more lines on the same payroll. Lines per person hour moves from 100 to about 105.

Every number in that paragraph is an illustration using round figures, not a benchmark and not a promise. The structure of the calculation is the point. Put your own pickers, your own shift length, your own measured travel share and your own pick rate into it, and you will have something defensible to take to whoever signs off the project.

The counter-force nobody nets off: replenishment

A location close to the packing area is valuable, which means there are not many of them, which means they tend to be small. That is the whole trade and it is the one a naive slotting exercise walks straight past.

Work an example through. An item is picked 200 times a day. Today it sits in a deep bulk location a long walk from packing, holding around 20 days of cover, so it gets replenished about once every 20 days. Move it to a small fast-pick location close to packing that holds about one day of cover, and every one of those 200 daily picks now starts closer to where it finishes.

It also now needs replenishing every day instead of once every 20 days. That is 19 extra replenishment events across those 20 days, and each one is a round trip to the reserve location and back, moving a pallet, which is slower and more equipment-dependent than a picker walking.

So the honest comparison is not picks against picks. It is picking time saved against replenishment time created, in a unit that accounts for the fact that a replenishment trip and a pick trip are not the same kind of work. Do that comparison per item and you get a ranked list where some obvious-looking moves turn out to be negative. Skip it and you get a plan that looks excellent on a slide and disappoints on the floor.

This is also why the answer is rarely to put the single fastest mover in the single closest location. The right answer usually gives close locations to items that are picked often and consume little space, because those are the ones that earn the position without demanding constant topping up.

Which is why "put the fast movers closest" is wrong on its own

The rule everyone knows is to put fast movers near packing. It is directionally right and it is incomplete, because it uses one variable where the decision has two.

A close location has a size. An item that is picked constantly but arrives on pallets and occupies a large volume will consume a close location that could instead have held several small, frequently picked items. The question is not which item is picked most. It is which item returns the most picking benefit per unit of close space it occupies.

That ratio is the reasoning behind cube-per-order index, a slotting method that dates to the 1960s and remains a sound starting point. Rank each product by physical volume divided by how often it is ordered; the small and frequently ordered ones belong closest to packing. It is not the last word, because it still says nothing about affinity, but it is a considerably better first word than frequency alone.

This is the practical reason item cube and weight are worth chasing down before an analysis. Without them, every recommendation is implicitly assuming that all items take up the same amount of space, which is never true and is most wrong exactly where the stakes are highest.

The part that has to be true first

Most warehouse optimisation is sold on top of location data nobody trusts.

If your system does not know where things actually are, optimising where they should be is theatre. A placement analysis run on inaccurate location data does not fail loudly. It produces confident, specific, wrong answers, and it produces them in a format that looks authoritative enough to act on.

Location data goes wrong for an ordinary reason: updating it is harder than not updating it. If recording a move takes several steps in a system people find slow, those moves stop being recorded, and the data drifts from reality one unrecorded pallet at a time. Nobody decided to stop maintaining it. The friction decided.

There is a cheap test before anyone commits to anything. Take a sample of locations, walk the floor, and compare what is physically there against what the system says. If the two agree, proceed. If they do not, that gap is the project, the slotting work waits, and any analysis run in the meantime will be precise about a building that does not exist.

We would rather say this before an engagement than during one, and it costs us work.

You do not have to move everything

The word re-slot makes people imagine emptying the building over a long weekend. That is not what a sensible project looks like.

In most operations a small proportion of products accounts for the majority of picking activity. Moving those captures most of the available benefit. Recommendations should always be ranked by value, so that you can stop at whatever point the effort stops being worth it, and so that the first week of work is the week with the largest return in it.

This also makes the project reversible in practice. A ranked list of moves is a list you can take the top twenty of, execute, measure, and then decide whether to continue. A whole-building re-slot is a decision you only get to make once, and it is the version that turns into a story people tell about the year the warehouse ground to a halt.

How these projects actually fail

Not from bad mathematics. The arithmetic is the easy part. They fail in recognisable ways, and all of them are avoidable.

• The location data was wrong and nobody checked. Covered above, and it is the most common one by a distance.
• The history window was too short. A quarter of data cannot see a season. An analysis built on a summer quarter will confidently move winter stock to the back of the building.
• Replenishment was never netted off. The picking saving was real and the replenishment cost it created was larger.
• There was no baseline. Nobody measured the before state, so six months later there is no way to settle whether it worked, and the argument gets decided by whoever is most senior in the room.
• Too much was moved at once. The disruption landed in a single window, the floor lost confidence, and the remaining phases were quietly cancelled.
• The pick faces could not hold the recommendation. A plan that ignores the physical capacity of the locations it is assigning to produces moves that cannot be executed, and it loses the floor’s trust the first time somebody tries one.
• Nobody owned it. Picking productivity sits with operations, the order data sits with IT, and the layout sits with whoever built the warehouse. The work lives in the gap between three departments and gaps do not have calendars.

What slotting will not fix

Worth saying plainly, because a page arguing for something should be able to say where it stops.

If your constraint is not picking, slotting will not help. An operation whose orders are waiting at packing or on the dock has a bottleneck somewhere else, and making picking faster will move the queue rather than shorten it. Find the constraint before optimising anything.

If the building itself is wrong for the work, slotting will only recover so much. A layout with a single narrow aisle to the packing area, or racking whose geometry forces long detours, has a physical ceiling that placement cannot rise above.

And if your demand has no pattern in it, there is less to find. Slotting works because past orders predict future ones. Where an item mix turns over almost completely each season, the analysis has to be rerun far more often for a smaller reward each time, which is a real answer and sometimes the answer is that this is not your best available project.

What you need before anyone can analyse this

Less than people expect, and none of it requires access to a live system.

Twelve months of order lines, meaning order number, item, quantity and date. A location master showing which item is stored where. Item cube and weight are a considerable help where they exist and are not a blocker where they do not.

The order number field is doing more work than it looks. It is what makes affinity computable, because it is the only thing that records which lines travelled together. An extract of item movements without the order number will support a velocity analysis and cannot support an affinity one.

Twelve months rather than a quarter, and this matters more than it sounds. Ninety days cannot see a seasonal cycle. An analysis built on a summer quarter will confidently recommend moving your winter stock into the back of the building, and it will be wrong in a way that is expensive and slow to notice.

All of these are standard exports from any ERP or warehouse management system. No production access is required, and nothing needs to be installed.

How to know whether it worked

Measure before you move. If you cannot say what the current state costs, you cannot prove you improved it, and you will end up arguing about whether the project worked using numbers nobody agrees on.

The honest method is to reconstruct the trips implied by real historical orders against the current layout, then reconstruct the same orders against the proposed layout. The difference is the waste you are removing, and it can be expressed in distance, in time, or in labour cost.

Be clear about what that method is and is not. It is a model. It does not know that a picker took a shortcut through a cross aisle, or waited for a forklift, or picked two orders at once because it was quicker. What makes it useful is not that it reproduces reality, but that it applies exactly the same assumptions to the before case and the after case, so the difference between them is meaningful even where the absolute numbers are not. Anyone presenting the absolute figure as measured truth is overselling the method.

Then hold the result to lines picked and shipped per person hour once the moves are done, over a period long enough for volume and mix to be comparable to the baseline. A quiet fortnight after a busy one will show an improvement that has nothing to do with the project.

A slotting project that cannot be checked afterwards was a rearrangement, not an optimisation.

For the full ground-up treatment of the problem this sits inside, read Warehouse Pickup Optimization: The Operator’s Guide.

If you run Dynamics 365 and want to know what the ERP does and does not do here, see Can Dynamics 365 optimize inventory placement?.